Why we say soul, not data

The first time we mounted a .laplaspack and asked a question, the AI answered with the author's own reasoning. Not a summary of it. The actual chain — what he tried first, why he abandoned it, what the abandonment taught him, and how that led to the decision he eventually made. His why, in order.

It did not feel like a database query.

It felt like someone walked into the room.

We sat there for a while. Then we stopped calling them files.

What a soul technically is

Strip the word back and it's an engineering object. A soul is a whole mind in one file: the entities someone thought about, the causal reasoning DAG that connects them — this decision because of that failure, this belief revised by that result — and the full history of how all of it evolved. The file is Ed25519-sealed and verifiable offline. No server has to vouch for it. The signature is the provenance.

Double-click it and it mounts. From that moment, your connected AI doesn't think about that person's notes. It thinks with their mind — their definitions, their dead ends, their reasons, in the order they actually happened.

Why "data" was never the right word

Data is what's left after the thinking evaporates. A spreadsheet of conclusions with the reasoning rinsed out. You can query data. You can't ask data why.

This is, quietly, the oldest loss in human work. We have only ever shipped results. The paper, not the four dead ends that shaped it. The codebase, not the argument that settled its architecture. The company, not the nights the founder almost chose differently. The process — the actual thinking — evaporated with the person, every single time.

Souls ship the process. For the first time, the why survives the work.

That's why the word matters. "Data" describes what a thing is made of. "Soul" describes what it does when it enters a room: it carries a particular person's way of reasoning into a mind that isn't theirs. The first word undersells it. The second one is simply accurate.